4. 1029 title pg 7/10/02 2:45 PM Page 1
CRC PR ESS
Boca Raton London New York Washington, D.C.
TEXTBOOK
of
RECEPTOR
PHARMACOLOGY
Second Edition
Edited by
John C. Foreman, D.Sc., F.R.C.P.
Department of Pharmacology
University College London
United Kingdom
Torben Johansen, M.D.
Department of Physiology and Pharmacology
University of Southern Denmark
Denmark
6. Preface
For about four decades now, a course in receptor pharmacology has been given at University College
London for undergraduate students in their final year of study for the Bachelor of Science degree
in pharmacology. More recently, the course has also been taken by students reading for the Bachelor
of Science degree in medicinal chemistry. The students following the course have relied for their
reading upon a variety of sources, including original papers, reviews, and various textbooks, but
no single text brought together the material included in the course. Also, almost continuously since
1993, we have organized courses for graduate students and research workers from the pharmaceu-
tical industry from the Nordic and European countries. In many cases, generous financial support
from the Danish Research Academy and the Nordic Research Academy has made this possible.
These courses, too, were based on those for students at University College London, and we are
grateful for the constructive criticisms of the many students on all of the courses that have shaped
this book.
The first edition of the book provided a single text for the students, and the enthusiasm with
which it was received encouraged us to work on a second edition. There have been very significant
steps forward since the first edition of this book, particularly in the molecular biology of receptors.
These advances are reflected in the rewritten chapters for the section of this book that deals with
molecular biology. At the same time, we realized that in the first edition we included too much
material that was distant from the receptors themselves. To include all the cellular biology that is
consequent upon a receptor activation is really beyond the scope of any book. Hence, we have
omitted from the second edition the material on intracellular second messengers such as calcium,
the cyclic nucleotides, and phospholipids. The second edition now concentrates on cell membrane
receptors themselves, together with their immediate signal transducers: ion channels, heterotrimeric
G-proteins, and tyrosine kinases.
The writers of the chapters in this book have been actively involved in teaching the various
courses, and our joint aim has been to provide a logical introduction to the study of drug receptors.
Characterization of drug receptors involves a number of different approaches: quantitative descrip-
tion of the functional studies with agonists and antagonists, quantitative description of the binding
of ligands to receptors, the molecular structure of drug receptors, and the elements that transduce
the signal from the activated receptor to the intracellular compartment.
The book is intended as an introductory text on receptor pharmacology but further reading has
been provided for those who want to follow up on topics. Some problems are also provided for
readers to test their grasp of material in some of the chapters.
John C. Foreman
Torben Johansen
7.
8. The Editors
John C. Foreman, B.Sc., Ph.D., D.Sc., M.B., B.S., F.R.C.P., is Professor of Immunopharmacology
at University College London. He has also been a Visiting Professor at the University of Southern
Denmark, Odense, Denmark, and the University of Tasmania, Hobart, Australia. Dr. Foreman is
Dean of Students at University College London and also Vice-Dean of the Faculty of Life Sciences.
He was Senior Tutor of University College London from 1989 to 1996 and Admissions Tutor for
Medicine from 1982 to 1993. Dr. Foreman was made a Fellow of University College London in
1993 and received the degree of Doctor of Science from the University of London in the same
year. He was elected to the Fellowship of the Royal College of Physicians in 2001. Dr. Foreman
initially read medicine at University College London but interrupted his studies in medicine to take
the B.Sc. and Ph.D. in pharmacology before returning to complete the medical degrees, M.B., B.S.,
which he obtained in 1976. After internships at Peterborough District Hospital, he spent two years
as Visiting Instructor of Medicine, Division of Clinical Immunology, Johns Hopkins University
Schools of Medicine, Baltimore, MD. He then returned to University College London, where he
has remained on the permanent staff.
Dr. Foreman is a member of the British Pharmacological Society and the Physiological Society
and served as an editor of the British Journal of Pharmacology from 1980 to 1987 and again from
1997 to 2000. He has been an associate editor of Immunopharmacology and is a member of the
editorial boards of Inflammation Research and Pharmacology and Toxicology. Dr. Foreman has
presented over 70 invited lectures around the world. He is co-editor of the Textbook of Immuno-
pharmacology, now in its third edition, and has published approximately 170 research papers, as
well as reviews and contributions to books. His current major research interests include bradykinin
receptors in the human nasal airway, mechanisms of activation of dendritic cells, and the control
of microvascular circulation in human skin.
Torben Johansen, M.D., dr. med., is Docent of Pharmacology, Department of Physiology and
Pharmacology, Institute of Medical Biology, Faculty of Health Sciences, University of Southern
Denmark. Dr. Johansen obtained his M.D. degree in 1970 from the University of Copenhagen,
became a research fellow in the Department of Pharmacology of Odense University in 1970, lecturer
in 1972, and senior lecturer in 1974. Since 1990, he has been Docent of Pharmacology. In 1979, he
was a visiting research fellow for three months at the University Department of Clinical Pharma-
cology, Oxford University, and in 1998 and 2001 he was a visiting research fellow at the Department
of Pharmacology, University College London. In 1980, he did his internship in medicine and surgery
at Odense University Hospital. He obtained his Dr. Med. Sci. in 1988 from Odense University.
Dr. Johansen is a member of the British Pharmacological Society, the Physiological Society, the
Scandinavian Society for Physiology, the Danish Medical Association, the Danish Pharmacological
Society, the Danish Society for Clinical Pharmacology, and the Danish Society for Hypertension.
He has published 70 research papers in refereed journals. His current major research interests are
NMDA receptors in the substantia nigra in relation to cell death in Parkinson’s disease and also ion
transport and signaling in mast cells in relation to intracellular pH and volume regulation.
9.
10. Contributors
Sir James W. Black, Nobel Laureate,
F.R.S.
James Black Foundation
London, United Kingdom
David A. Brown, F.R.S.
Department of Pharmacology
University College London
London, United Kingdom
Jan Egebjerg, Ph.D.
Department for Molecular and Structural
Biology
Aarhus University
Aarhus, Denmark
Steen Gammeltoft, M.D.
Department of Clinical Biochemistry
Glostrup Hospital
Glostrup, Denmark
Alasdair J. Gibb, Ph.D.
Department of Pharmacology
University College London
London, United Kingdom
Dennis G. Haylett, Ph.D.
Department of Pharmacology
University College London
London, United Kingdom
Birgitte Holst
Department of Pharmacology
University of Copenhagen
Panum Institute
Copenhagen, Denmark
Donald H. Jenkinson, Ph.D.
Department of Pharmacology
University College London
London, United Kingdom
IJsbrand Kramer, Ph.D.
Section of Molecular and Cellular Biology
European Institute of Chemistry and Biology
University of Bordeaux 1
Talence, France
Thue W. Schwartz, M.D.
Department of Pharmacology
University of Copenhagen
Panum Institute
Copenhagen, Denmark
11.
12. Contents
Section I: Drug–Receptor Interactions
Chapter 1
Classical Approaches to the Study of Drug–Receptor Interactions..................................................3
Donald H. Jenkinson
Section II: Molecular Structure of Receptors
Chapter 2
Molecular Structure and Function of 7TM G-Protein-Coupled Receptors....................................81
Thue W. Schwartz and Birgitte Holst
Chapter 3
The Structure of Ligand-Gated Ion Channels...............................................................................111
Jan Egebjerg
Chapter 4
Molecular Structure of Receptor Tyrosine Kinases ......................................................................131
Steen Gammeltoft
Section III: Ligand-Binding Studies of Receptor
Chapter 5
Direct Measurement of Drug Binding to Receptors.....................................................................153
Dennis G. Haylett
Section IV: Transduction of the Receptor Signal
Chapter 6
Receptors Linked to Ion Channels: Mechanisms of Activation and Block..................................183
Alasdair J. Gibb
Chapter 7
G-Proteins.......................................................................................................................................213
David A. Brown
13. Chapter 8
Signal Transduction through Protein Tyrosine Kinases................................................................237
IJsbrand Kramer
Section V: Receptors as Pharmaceutical Targets
Chapter 9
Receptors as Pharmaceutical Targets.............................................................................................271
James W. Black
Index ..............................................................................................................................................279
17. 4 Textbook of Receptor Pharmacology, Second Edition
1.5 Inhibitory Actions at Receptors: I. Surmountable Antagonism ............................................41
1.5.1 Overview of Drug Antagonism..................................................................................41
1.5.1.1 Mechanisms Not Involving the Agonist Receptor Macromolecule...........41
1.5.1.2 Mechanisms Involving the Agonist Receptor Macromolecule ..................42
1.5.2 Reversible Competitive Antagonism..........................................................................43
1.5.3 Practical Applications of the Study of Reversible Competitive Antagonism ...........47
1.5.4 Complications in the Study of Reversible Competitive Antagonism........................49
1.5.5 Appendix to Section 1.5: Application of the Law of Mass Action to Reversible
Competitive Antagonism ............................................................................................52
1.6 Inhibitory Actions at Receptors: II. Insurmountable Antagonism ........................................53
1.6.1 Irreversible Competitive Antagonism.........................................................................53
1.6.2 Some Applications of Irreversible Antagonists .........................................................54
1.6.2.1 Labeling Receptors .....................................................................................54
1.6.2.2 Counting Receptors.....................................................................................55
1.6.2.3 Receptor Protection Experiments ...............................................................55
1.6.3 Effect of an Irreversible Competitive Antagonist on the Response to an
Agonist........................................................................................................................55
1.6.4 Can an Irreversible Competitive Antagonist Be Used to Find the Dissociation
Equilibrium Constant for an Agonist?.......................................................................57
1.6.5 Reversible Noncompetitive Antagonism....................................................................59
1.6.6 A More General Model for the Action of Agonists, Co-agonists, and
Antagonists .................................................................................................................63
1.6.7 Appendices to Section 1.6..........................................................................................64
1.6.7.1 Appendix 1.6A. A Note on the Term Allosteric ........................................64
1.6.7.2 Appendix 1.6B. Applying the Law of Mass Action to the Scheme of
Figure 1.28 ..................................................................................................66
1.7 Concluding Remarks ..............................................................................................................70
1.8 Problems .................................................................................................................................70
1.9 Further Reading......................................................................................................................71
1.10 Solutions to Problems ............................................................................................................72
1.1 INTRODUCTION
The term receptor is used in pharmacology to denote a class of cellular macromolecules that are
concerned specifically and directly with chemical signaling between and within cells. Combination
of a hormone, neurotransmitter, or intracellular messenger with its receptor(s) results in a change
in cellular activity. Hence, a receptor must not only recognize the particular molecules that activate
it, but also, when recognition occurs, alter cell function by causing, for example, a change in
membrane permeability or an alteration in gene transcription.
The concept has a long history. Mankind has always been intrigued by the remarkable ability
of animals to distinguish different substances by taste and smell. Writing in about 50 B.C., Lucretius
(in De Rerum Natura, Liber IV) speculated that odors might be conveyed by tiny, invisible “seeds”
with distinctive shapes which would have to fit into minute “spaces and passages” in the palate
and nostrils. In his words:
Some of these must be smaller, some greater, they must be three-cornered for some creatures, square
for others, many round again, and some of many angles in many ways.
The same principle of complementarity between substances and their recognition sites is
implicit in John Locke’s prediction in his Essay Concerning Human Understanding (1690):
18. Classical Approaches to the Study of Drug–Receptor Interactions 5
Did we but know the mechanical affections of the particles of rhubarb, hemlock, opium and a man, as
a watchmaker does those of a watch, … we should be able to tell beforehand that rhubarb will purge,
hemlock kill and opium make a man sleep.
(Here, mechanical affections could be replaced in today’s usage by chemical affinities.)
Prescient as they were, these early ideas could only be taken further when, in the early 19th
century, it became possible to separate and purify the individual components of materials of plant
and animal origin. The simple but powerful technique of fractional crystallization allowed plant
alkaloids such as nicotine, atropine, pilocarpine, strychnine, and morphine to be obtained in a pure
form for the first time. The impact on biology was immediate and far reaching, for these substances
proved to be invaluable tools for the unraveling of physiological function. To take a single example,
J. N. Langley made great use of the ability of nicotine to first activate and then block nerves
originating in the autonomic ganglia. This allowed him to map out the distribution and divisions
of the autonomic nervous system.
Langley also studied the actions of atropine and pilocarpine, and in 1878 he published (in the
first volume of the Journal of Physiology, which he founded) an account of the interactions between
pilocarpine (which causes salivation) and atropine (which blocks this action of pilocarpine). Con-
firming and extending the pioneering work of Heidenhain and Luchsinger, Langley showed that
the inhibitory action of atropine could be overcome by increasing the dose of pilocarpine. Moreover,
the restored response to pilocarpine could in turn be abolished by further atropine. Commenting
on these results, Langley wrote:
We may, I think, without too much rashness, assume that there is some substance or substances in the
nerve endings or [salivary] gland cells with which both atropine and pilocarpine are capable of forming
compounds. On this assumption, then, the atropine or pilocarpine compounds are formed according to
some law of which their relative mass and chemical affinity for the substance are factors.
If we replace mass by concentration, the second sentence can serve as well today as when it
was written, though the nature of the law which Langley had inferred must exist was not to be
formulated (in a pharmacological context) until almost 60 years later. It is considered in Section
1.5.2 below.
J. N. Langley maintained an interest in the action of plant alkaloids throughout his life. Through
his work with nicotine (which can contract skeletal muscle) and curare (which abolishes this action
of nicotine and also blocks the response of the muscle to nerve stimulation, as first shown by
Claude Bernard), he was able to infer in 1905 that the muscle must possess a “receptive substance”:
Since in the normal state both nicotine and curari abolish the effect of nerve stimulation, but do not
prevent contraction from being obtained by direct stimulation of the muscle or by a further adequate
injection of nicotine, it may be inferred that neither the poison nor the nervous impulse acts directly
on the contractile substance of the muscle but on some accessory substance.
Since this accessory substance is the recipient of stimuli which it transfers to the contractile material,
we may speak of it as the receptive substance of the muscle.
At the same time, Paul Ehrlich, working in Frankfurt, was reaching similar conclusions, though
from evidence of quite a different kind. He was the first to make a thorough and systematic study
of the relationship between the chemical structure of organic molecules and their biological actions.
This was put to good use in collaboration with the organic chemist A. Bertheim. Together, they
prepared and tested more than 600 organometallic compounds incorporating mercury and arsenic.
Among the outcomes was the introduction into medicine of drugs such as salvarsan that were toxic
to pathogenic microorganisms responsible for syphilis, for example, at doses that had relatively
minor side effects in humans. Ehrlich also investigated the selective staining of cells by dyes, as
19. 6 Textbook of Receptor Pharmacology, Second Edition
well as the remarkably powerful and specific actions of bacterial toxins. All these studies convinced
him that biologically active molecules had to become bound in order to be effective, and after the
fashion of the time he expressed this neatly in Latin:
Corpora non agunt nisi fixata.*
In Ehrlich’s words (Collected Papers, Vol. III, Chemotherapy):
When the poisons and the organs sensitive to it do not come into contact, or when sensitiveness of the
organs does not exist, there can be no action.
If we assume that those peculiarities of the toxin which cause their distribution are localized in a special
group of the toxin molecules and the power of the organs and tissues to react with the toxin are localized
in a special group of the protoplasm, we arrive at the basis of my side chain theory. The distributive
groups of the toxin I call the “haptophore group” and the corresponding chemical organs of the
protoplasm the ‘receptor.’ … Toxic actions can only occur when receptors fitted to anchor the toxins
are present.
Today, it is accepted that Langley and Ehrlich deserve comparable recognition for the intro-
duction of the receptor concept. In the same years, biochemists studying the relationship between
substrate concentration and enzyme velocity had also come to think that enzyme molecules must
possess an “active site” that discriminates among various substrates and inhibitors.As often happens,
different strands of evidence had converged to point to a single conclusion.
Finally, a note on the two senses in which present-day pharmacologists and biochemists use
the term receptor. The first sense, as in the opening sentences of this section, is in reference to the
whole receptor macromolecule that carries the binding site for the agonist. This usage has become
common as the techniques of molecular biology have revealed the amino-acid sequences of more
and more signaling macromolecules. But, pharmacologists still sometimes employ the term receptor
when they have in mind only the particular regions of the macromolecule that are concerned in the
binding of agonist and antagonist molecules. Hence, receptor occupancy is often used as convenient
shorthand for the fraction of the binding sites occupied by a ligand.**
1.2 MODELING THE RELATIONSHIP BETWEEN AGONIST
CONCENTRATION AND TISSUE RESPONSE
With the concept of the receptor established, pharmacologists turned their attention to understanding
the quantitative relationship between drug concentration and the response of a tissue. This entailed,
first, finding out how the fraction of binding sites occupied and activated by agonist molecules
varies with agonist concentration, and, second, understanding the dependence of the magnitude of
the observed response on the extent of receptor activation.
Today, the first question can sometimes be studied directly using techniques that are described
in later chapters, but this was not an option for the early pharmacologists. Also, the only responses
that could then be measured (e.g., the contraction of an intact piece of smooth muscle or a change
in the rate of the heart beat) were indirect, in the sense that many cellular events lay between the
initial step (activation of the receptors) and the observed response. For these reasons, the early
workers had no choice but to devise ingenious indirect approaches, several of which are still
important. These are based on “modeling” (i.e., making particular assumptions about) the two
* Literally: entities do not act unless attached.
** Ligand means here a small molecule that binds to a specific site (or sites) on a receptor macromolecule. The term drug
is often used in this context, especially in the older literature.
20. Classical Approaches to the Study of Drug–Receptor Interactions 7
relationships identified above and then comparing the predictions of the models with the actual
behavior of isolated tissues. This will now be illustrated.
1.2.1 THE RELATIONSHIP BETWEEN LIGAND CONCENTRATION AND RECEPTOR
OCCUPANCY
We begin with the simplest possible representation of the combination of a ligand, A, with its
binding site on a receptor, R:
(1.1)
Here, binding is regarded as a bimolecular reaction and k+1 and k–1 are, respectively, the association
rate constant (M–1 s–1) and the dissociation rate constant (s–1).
The law of mass action states that the rate of a reaction is proportional to the product of the
concentrations of the reactants. We will apply it to this simple scheme, making the assumption that
equilibrium has been reached so that the rate at which AR is formed from A and R is equal to the
rate at which AR dissociates. This gives:
k+1[A][R] = k–1[AR]
where [R] and [AR] denote the concentrations of receptors in which the binding sites for A are
free and occupied, respectively.
It may seem odd to refer to receptor concentrations in this context when receptors can often
move only in the plane of the membrane (and even then perhaps to no more than a limited extent,
as many kinds of receptors are anchored). However, the model can be formulated equally well in
terms of the proportions of a population of binding sites that are either free or occupied by a ligand.
If we define pR as the proportion free,* equal to [R]/[R]T, where [R]T represents the total concen-
tration of receptors, and pAR as [AR]/[R]T, we have:
k+1[A]pR = k–1pAR
Because for now we are concerned only with equilibrium conditions and not with the rate at
which equilibrium is reached, we can combine k+1 and k–1 to form a new constant, KA = k–1/k+1,
which has the unit of concentration. KA is a dissociation equilibrium constant (see Appendix 1.2A
[Section 1.2.4.1]), though this is often abbreviated to either equilibrium constant or dissociation
constant. Replacing k+1 and k–1 gives:
[A]pR = KApAR
Because the binding site is either free or occupied, we can write:
pR + pAR = 1
Substituting for pR:
* pR can be also be defined as NR /N, where NR is the number of receptors in which the binding sites are free of A and N
is their total number. Similarly, pAR is given by NAR/N, where NAR is the number of receptors in which the binding site is
occupied by A. These definitions are used when discussing the action of irreversible antagonists (see Section 1.6.4).
A R AR
+
−
+
k
k
1
1
21. 8 Textbook of Receptor Pharmacology, Second Edition
Hence,*
(1.2)
This is the important Hill–Langmuir equation. A. V. Hill was the first (in 1909) to apply the law
of mass action to the relationship between ligand concentration and receptor occupancy at equi-
librium and to the rate at which this equilibrium is approached.** The physical chemist I. Langmuir
showed a few years later that a similar equation (the Langmuir adsorption isotherm) applies to the
adsorption of gases at a surface (e.g., of a metal or of charcoal).
In deriving Eq. (1.2), we have assumed that the concentration of A does not change as ligand
receptor complexes are formed. In effect, the ligand is considered to be present in such excess that
it is scarcely depleted by the combination of a little of it with the receptors, thus [A] can be regarded
as constant.
The relationship between pAR and [A] predicted by Eq. (1.2) is illustrated in Figure 1.1. The
concentration of A has been plotted using a linear (left) and a logarithmic scale (right). The value
of KA has been taken to be 1 µM. Note from Eq. (1.2) that when [A] = KA, pAR = 0.5; that is, half
of the receptors are occupied.
With the logarithmic scale, the slope of the line initially increases. The curve has the form of
an elongated S and is said to be sigmoidal. In contrast, with a linear (arithmetic) scale for [A],
sigmoidicity is not observed; the slope declines as [A] increases, and the curve forms part of a
rectangular hyperbola.
* If you find this difficult, see Appendix 1.2B at the end of this section.
** Hill had been an undergraduate student in the Department of Physiology at Cambridge where J. N. Langley suggested
to him that this would be useful to examine in relation to finding whether the rate at which an agonist acts on an isolated
tissue is determined by diffusion of the agonist or by its combination with the receptor.
FIGURE 1.1 The relationship between binding-site occupancy and ligand concentration ([A]; linear scale,
left; log scale, right), as predicted by the Hill–Langmuir equation. KA has been taken to be 1 µM for both curves.
K
p p
A
AR AR
A
[ ]
+ = 1
p
K
AR
A
A
A
=
+
[ ]
[ ]
22. Classical Approaches to the Study of Drug–Receptor Interactions 9
Equation (1.2) can be rearranged to:
Taking logs, we have:
Hence, a plot of log (pAR /(1 – pAR)) against log [A] should give a straight line with a slope of one.
Such a graph is described as a Hill plot, again after A. V. Hill, who was the first to employ it, and
it is often used when pAR is measured directly with a radiolabeled ligand (see Chapter 5). In practice,
the slope of the line is not always unity, or even constant, as will be discussed. It is referred to as
the Hill coefficient (nH); the term Hill slope is also used.
1.2.2 THE RELATIONSHIP BETWEEN RECEPTOR OCCUPANCY AND TISSUE RESPONSE
This is the second of the two questions identified at the start of Section 1.2, where it was noted
that the earliest pharmacologists had no choice but to use indirect methods in their attempts to
account for the relationship between the concentration of a drug and the tissue response that it
elicits. In the absence at that time of any means of obtaining direct evidence on the point, A. V.
Hill and A. J. Clark explored the consequences of assuming: (1) that the law of mass action applies,
so that Eq. (1.2), derived above, holds; and (2) that the response of the tissue is linearly related to
receptor occupancy. Clark went further and made the tentative assumption that the relationship
might be one of direct proportionality (though he was well aware that this was almost certainly an
oversimplification, as we now know it usually is).
Should there be direct proportionality, and using y to denote the response of a tissue (expressed
as a percentage of the maximum response attainable with a large concentration of the agonist), the
relationship between occupancy* and response becomes:
(1.3)
Combining this with Eq. (1.2) gives an expression that predicts the relationship between the
concentration of the agonist and the response that it elicits:
(1.4)
This is often rearranged to:
(1.5)
* Note that no distinction is made here between occupied and activated receptors; it is tacitly assumed that all the receptors
occupied by agonist molecules are in an active state, hence contributing to the initiation of the tissue response that is
observed. As we shall see in the following sections, this is a crucial oversimplification.
p
p K
AR
AR A
A
1−
=
[ ]
log log[ ] log
p
p
K
AR
AR
A
A
1−
⎛
⎝
⎜
⎞
⎠
⎟ = −
y
p
100
= AR
y
K
100
=
+
[ ]
[ ]
A
A
A
y
y K
100 −
=
[ ]
A
A
23. 10 Textbook of Receptor Pharmacology, Second Edition
Taking logs,
The applicability of this expression (and by implication Eq. (1.4)) can be tested by measuring
a series of responses (y) to different concentrations of A and then plotting log (y/(100 – y)) against
log [A] (the Hill plot). If Equation (1.4) holds, a straight line with a slope of 1 should be obtained.
Also, were the underlying assumptions to be correct, the value of the intercept of the line on the
abscissa (i.e., when the response is half maximal) would give an estimate of KA. A. J. Clark was
the first to test this using the responses of isolated tissues, and Figure 1.2 illustrates some of his
results. Figure 1.2A shows that Eq. (1.4) provides a reasonably good fit to the experimental values.
Also, the slopes of the Hill plots in Figure 1.2B are close to unity (0.9 for the frog ventricle, 0.8
for the rectus abdominis). While these findings are in keeping with the simple model that has been
outlined, they do not amount to proof that it is correct. Indeed, later studies with a wide range of
tissues have shown that many concentration–response relationships cannot be fitted by Eq. (1.4).
For example, the Hill coefficient is almost always greater than unity for responses mediated by
ligand-gated ion channels (see Appendix 1.2C [Section 1.2.4.3] and Chapter 6). What is more, it
is now known that with many tissues the maximal response (for example, contraction of intestinal
smooth muscle) can occur when an agonist such as acetylcholine occupies less than a tenth of the
available receptors, rather than all of them as postulated in Eq. (1.3). By the same token, when an
agonist is applied at the concentration (usually termed the [A]50 or EC50) required to produce a
half-maximal response, receptor occupancy may be as little as 1% in some tissues,* rather than
the 50% expected if the response is directly proportional to occupancy. An additional complication
is that many tissues contain enzymes (e.g., cholinesterase) or uptake processes (e.g., for noradren-
aline) for which agonists are substrates. Because of this, the agonist concentration in the inner
regions of an isolated tissue may be much less than in the external solution.
Pharmacologists have therefore had to abandon (sometimes rather reluctantly and belatedly)
not only their attempts to explain the shapes of the dose–response curves of complex tissues in
terms of the simple models first explored by Clark and by Hill, but also the hope that the value of
the concentration of an agonist that gives a half-maximal response might provide even an approx-
imate estimate of KA. Nevertheless, as Clark’s work showed, the relationship between the concen-
tration of an agonist and the response of a tissue commonly has the same general form shown in
Figure 1.1. In keeping with this, concentration–response curves can often be described empirically,
and at least to a first approximation, by the simple expression:
(1.6)
This is usually described as the Hill equation (see also Appendix 1.2C [Section 1.2.4.3]). Here,
nH is again the Hill coefficient, and y and ymax are, respectively, the observed response and the
maximum response to a large concentration of the agonist, A. [A]50 is the concentration of A at
which y is half maximal. Because it is a constant for a given concentration–response relationship,
it is sometimes denoted by K. While this is algebraically neater (and was the symbol used by Hill),
it should be remembered that K in this context does not necessarily correspond to an equilibrium
constant. Employing [A]50 rather than K in Eq. (1.6) helps to remind us that the relationship between
* For evidence on this, see Section 1.6 on irreversible antagonists.
log log[ ] log
y
y
K
100 −
⎛
⎝
⎜
⎞
⎠
⎟ = −
A A
y y
n
n n
= max
[ ]
[ ] [ ]
A
A A
H
H H
50 +
24. Classical Approaches to the Study of Drug–Receptor Interactions 11
FIGURE 1.2 (Upper) Concentration–response relationship for the action of acetylcholine in causing contrac-
tion of the frog rectus abdominis muscle. The curve has been drawn using Eq. (1.4). (Lower) Hill plots for
the action of acetylcholine on frog ventricle (curve I) and rectus abdominis (curve II). (From Clark, A. J., J.
Physiol., 61, 530–547, 1926.)
25. 12 Textbook of Receptor Pharmacology, Second Edition
[A] and response is here being described rather than explained in terms of a model of receptor
action. This is an important difference.
1.2.3 THE DISTINCTION BETWEEN AGONIST BINDING AND RECEPTOR ACTIVATION
Finally, we return to models of receptor action and to a further limitation of the early attempts to
account for the shapes of concentration–response curves. As already noted, the simple concepts
expressed in Eqs. (1.3) and (1.4) do not distinguish between the occupation and the activation of
a receptor by an agonist. This distinction, it is now appreciated, is crucial to the understanding of
the action of agonists and partial agonists. Indeed all contemporary accounts of receptor activation
take as their starting point a mechanism of the following kind:*
(1.7)
Here, the occupied receptors can exist in two forms, one of which is inactive (AR) and the other
active (AR*) in the sense that its formation leads to a tissue response. AR and AR* can interconvert
(often described as isomerization), and at equilibrium the receptors will be distributed among the
R, AR, and AR* conditions.** The position of the equilibrium between AR and AR*, and hence
the magnitude of the maximum response of the tissue, will depend on the value of the equilibrium
constant E.*** Suppose that a very large concentration of the agonist A is applied, so that all the
binding sites are occupied (i.e., the receptors are in either the AR or the AR* state). If the position
of the equilibrium strongly favors AR, with few active (AR*) receptors, the response will be
relatively small. The reverse would apply for a very effective agonist. This will be explained in
greater detail in Sections 1.4.3–7, where we will also look into the relationship between agonist
concentration and the fraction of receptors in the active state.
1.2.4 APPENDICES TO SECTION 1.2
1.2.4.1 Appendix 1.2A: Equilibrium, Dissociation, and Affinity Constants
Confusingly, all of these terms are in current use to express the position of the equilibrium between
a ligand and its receptors. The choice arises because the ratio of the rate constants k–1 and k+1 can
be expressed either way up. In this chapter, we take KA to be k–1/k+1, and it is then strictly a
dissociation equilibrium constant, often abbreviated to either dissociation constant or equilibrium
constant. The inverse ratio, k+1/k–1, gives the association equilibrium constant, which is usually
referred to as the affinity constant.
One way to reduce the risk of confusion is to express ligand concentrations in terms of KA.
This “normalized” concentration is defined as [A]/KA and will be denoted here by the symbol ¢A.
We can therefore write the Hill–Langmuir equation in three different though equivalent ways:
where the terms are as follows:
* This will be described as the del Castillo–Katz scheme, as it was first applied to receptor action by J. del Castillo and B.
Katz (University College London) in 1957 (see also Section 1.4.3).
** The scheme is readily extended to include the possibility that some of the receptors may be active even in the absence
of agonist (see Section 1.4.7).
*** This constant is sometimes denoted by L or by K2. E has been chosen for this introductory account because of the
relation to efficacy and also because it is the term used in an important review by Colquhoun (1998) on binding, efficacy,
and the effects thereon of receptor mutations.
A + R
inactive
vacant KA
AR
inactive
occupied
E
AR *
active
occupied
p
K
K
K
AR
A
A
A
A
A
A
A
A
A
=
+
=
′
+ ′
=
+
[ ]
[ ]
[ ]
[ ]
¢
¢
1 1
26. Classical Approaches to the Study of Drug–Receptor Interactions 13
1.2.4.2 Appendix 1.2B: Step-by-Step Derivation of the Hill–Langmuir
Equation
We start with the two key equations given in Section 1.2.1:
[A]pR = KApAR (A.1)
pR + pAR = 1 (A.2)
From Eq. (A.1), (A.3)
Next, use Eq. (A.3) to replace pR in Eq. (A.2). This is done because we wish to find pAR:
The Hill–Langmuir equation may be rearranged by cross-multiplying:
pARKA + pAR[A] = [A]
pARKA = [A](1 – pAR)
Taking logs,
Abbreviation Unit
Dissociation equilibrium constant KA M
Affinity constant K′
A M–1
Normalized concentration ¢A —
p
K
p
R
A
AR
A
=
[ ]
K
p p
A
AR AR
[ ]
A
+ = 1
p
K
AR
A
[ ]
A
+
⎛
⎝
⎜
⎞
⎠
⎟ =
1 1
p
K
AR
A A
A
+
⎛
⎝
⎜
⎞
⎠
⎟ =
[ ]
[ ]
1
p
K A
AR
A
A]
=
+
[
[ ]
p
p K
AR
AR A
A]
1−
=
[
log log[ ] log
p
p
K
AR
AR
A
A
1−
⎛
⎝
⎜
⎞
⎠
⎟ = −
Remember, if ax = by, then x = (b/a)y.
Remember, ax + x = x(a + 1).
Remember, (s/t) + 1 = (s + t)/t.
Remember, if x(u/v) = 1, then x = (v/u).
For cross-multiplication, if (a/b) = (c/d), then
(a × d) = (c × b). Remember, y = x/(a + x) is
the same as (y/1) = x/(a + x), which is ready for
cross-multiplication.
Remember, log (a/b) = log a – log b.
27. 14 Textbook of Receptor Pharmacology, Second Edition
1.2.4.3 Appendix 1.2C: The Hill Equation and Hill Plot
In some of his earliest work, published in 1910, A. V. Hill examined how the binding of oxygen
to hemoglobin varied with the oxygen partial pressure. He found that the relationship between the
two could be fitted by the following equation:
Here, y is the fractional binding, x is the partial pressure of O2, K′ is an affinity constant, and n is
a number which in Hill’s work varied from 1.5 to 3.2.
This equation can also be written as:
(1.8a)
where Ke = 1/K′, and as:
(1.8b)
This final variant is convenient because K has the same units as x and, moreover, is the value of x
for which y is half maximal.
Eq. (1.8b) can be rearranged and expressed logarithmically as:
Hence, a Hill plot (see earlier discussion) should give a straight line of slope n.
Hill plots are often used in pharmacology, where y may be either the fractional response of a
tissue or the amount of a ligand bound to its binding site, expressed as a fraction of the maximum
binding, and x is the concentration. It is sometimes found (especially when tissue responses are
measured) that the Hill coefficient differs markedly from unity. What might this mean?
One of the earliest explanations to be considered was that n molecules of ligand might bind
simultaneously to a single binding site, R:
This would lead to the following expression for the proportion of binding sites occupied by A:
y
K x
K x
n
n
=
′
+ ′
1
y
x
K x
n
e
n
=
+
y
x
K x
n
n n
=
+
log log log
y
y
n x n K
1−
⎛
⎝
⎜
⎞
⎠
⎟ = −
n n
A R A R
+
p
K
n
n
n
A R
A
A
=
+
[ ]
[ ]
28. Classical Approaches to the Study of Drug–Receptor Interactions 15
where K is the dissociation equilibrium constant. Hence, the Hill plot would be a straight line with
a slope of n. However, this model is quite unlikely to apply. Extreme conditions aside, few examples
exist of chemical reactions in which three or more molecules (e.g., two of A and one of R) must
combine simultaneously. Another explanation has to be sought. One possibility arises when the
tissue response measured is indirect, in the sense that a sequence of cellular events links receptor
activation to the response that is finally observed. The Hill coefficient may not then be unity (or
even a constant) because of a nonlinear and variable relation between the proportion of receptors
activated and one or more of the events that follow.
Even when it is possible to observe receptor activation directly, the Hill coefficient may still
be found not to be unity. This has been studied in detail for ligand-gated ion channels such as the
nicotinic receptor for acetylcholine. Here the activity of individual receptors can be followed as it
occurs by measuring the tiny flows of electrical current through the ion channel intrinsic to the
receptor (see Section 1.4.3 and Chapter 6). On determining the relationship between this response
and agonist concentration, the Hill coefficient is observed to be greater than unity (characteristically
1.3–2) and to change with agonist concentration. The explanation is to be found in the structure
of this class of receptor. Each receptor macromolecule is composed of several (often five) subunits,
of which two carry binding sites for the agonist. Both of these sites must be occupied for the
receptor to become activated, at least in its normal mode. The scheme introduced in Section 1.2.3
must then be elaborated:
(1.9)
Suppose that the two sites are identical (an oversimplification) and that the binding of the first
molecule of agonist does not affect the affinity of the site that remains vacant. The dissociation
equilibrium constant for each site is denoted by KA and the equilibrium constant for the isomer-
ization between A2R and A2R* by E, so that [A2R*] = E[A2R].
The proportion of receptors in the active state (A2R*) is then given by:
(1.10)
This predicts a nonlinear Hill plot. Its slope will vary with [A] according to:
When [A] is small in relation to KA, nH approximates to 2. However, as [A] is increased, nH tends
toward unity.
On the same scheme, the amount of A that is bound (expressed as a fraction, pbound, of the
maximum binding when [A] is very large, so that all the sites are occupied) is given by:
(1.11)
A R AR A A R A R
+ + *
2 2
p
E
K E
A R
A
A
A A
2
2
2 2
*
[ ]
( [ ]) [ ]
=
+ +
n
K
K
A
H
A
A])
A]
=
+
+
2
2
( [
[
p
K E
K E
bound =
+ +
+ +
[ ]( [ ]) [ ]
( [ ]) [
A A A
A A]
A
A
2
2
2
29. 16 Textbook of Receptor Pharmacology, Second Edition
The Hill plot for binding would be nonlinear with a Hill coefficient given by:
(1.12)
This approximates to unity if [A] is either very large or very small. In between, nH may be as much
as 2 for very large values of E. It is noteworthy that this should be so even though the affinities
for the first and the second binding steps have been assumed to be the same, provided only that
some isomerization of the receptor to the active form occurs. This is because isomerization increases
the total amount of binding by displacing the equilibria shown in Eq. (1.9) to the right — that is,
toward the bound forms of the receptor.
We now consider what would happen if the binding of the first molecule of agonist altered the
affinity of the second identical site. The dissociation equilibrium constants for the first and second
bindings will be denoted by KA(1) and KA(2), respectively, and E is defined as before.
The proportion of receptors in the active state (A2R*) is then given by:
(1.13)
and the Hill coefficient nH would be:
These relationships are discussed further in Chapter 6 (see Eqs. (6.4) and (6.5)).
Using the same scheme, the amount of A that is bound is given by:
(1.14)
The Hill plot would again be nonlinear with the Hill coefficient given by:
(1.15)
This approximates to unity if [A] is either very large or very small. In between, nH may be greater
(up to 2) or less than 1, depending on the magnitude of E and on the relative values of KA(1) and
KA(2). If, for simplicity, we set E to 0 and if KA(2) < KA(1), then nH > 1, and there is said to be positive
cooperativity. Negative cooperativity occurs when KA(2) > KA(1) and nH is then < 1. This is discussed
further in Chapter 5 where plots of Eqs. (1.14) and (1.15) are shown (Figure 5.3) for widely ranging
values of the ratio of KA(1) to KA(2), and with E taken to be zero.
1.2.4.4 Appendix 1.2D: Logits, the Logistic Equation, and their Relation to
the Hill Plot and Equation
The logit transformation of a variable p is defined as:
n
K E K
K K E
H
A A
A A
A A A
A A
=
+ + +
+ + +
( [ ]) [ ]( [ ])
( [ ]){ ( )[ ]}
2
2
1
p
E
K K K E
A R
A A A
2
A
A A
*
( ) ( ) ( )
[ ]
[ ] ( )[ ]
=
+ + +
2
1 2 2
2
2 1
n
K
K
H
A
A
A
A
=
+
+
2
2
1
1
( [ ])
[ ]
( )
( )
p
K E
K K K E
bound =
+ +
+ + +
A
A A A
A A
A A
( )
( ) ( ) ( )
[ ] ( )[ ]
[ ] ( )[ ]
2
2
1 2 2
2
1
2 1
n
K K E K
K K E
H
A A A
A A
A A
A A
=
+ + +
+ + +
( ) ( ) ( )
( ) ( )
( )[ ]( [ ])
( [ ]){ ( )[ ]}
1 2 1
1 2
1 2
1
30. Classical Approaches to the Study of Drug–Receptor Interactions 17
Hence, the Hill plot can be regarded as a plot of logit (p) against the logarithm of concentration
(though it is more usual to employ logs to base 10 than to base e).
It is worth noting the distinction between the Hill equation and the logistic equation, which
was first formulated in the 19th century as a means of describing the time-course of population
increase. It is defined by the expression:
(1.16)
This is easily rearranged to:
Hence,
If we redefine a as –loge K, and x as loge z, then
(1.17)
which is a form of the Hill equation (see Eq. (1.8a)). However, note that Eq. (1.17) has been
obtained from Eq. (1.16) only by transforming one of the variables. It follows that the terms logistic
equation (or curve) and Hill equation (or curve) should not be regarded as interchangeable. To
illustrate the distinction, if the independent variable in each equation is set to zero, the dependent
variable becomes 1/(1 + e–a) in Eq. (1.16) as compared with zero in Eq. (1.17).
1.3 THE TIME COURSE OF CHANGES IN RECEPTOR OCCUPANCY
1.3.1 INTRODUCTION
At first glance, the simplest approach to determining how quickly a drug combines with its receptors
might seem to be to measure the rate at which it acts on an isolated tissue, but two immediate
problems arise. The first is that the exact relationship between the effect on a tissue and the
proportion of receptors occupied by the drug is often not known and cannot be assumed to be
simple, as we have already seen. A half-maximal tissue response only rarely corresponds to half-
maximal receptor occupation. We can take as an example the action of the neuromuscular blocking
agent tubocurarine on the contractions that result from stimulation of the motor nerve supply to
skeletal muscle in vitro. The rat phrenic nerve–diaphragm preparation is often used in such exper-
iments. Because neuromuscular transmission normally has a large safety margin, the contractile
response to nerve stimulation begins to fall only when tubocurarine has occupied on average more
than 80% of the binding sites on the nicotinic acetylcholine receptors located on the superficial
logit e
[ ] log
p
p
p
=
−
⎛
⎝
⎜
⎞
⎠
⎟
1
p
e a bx
=
+ − +
1
1 ( )
p
p
ea bx
1−
= +
logit[ ] log
p
p
p
a bx
e
=
−
⎛
⎝
⎜
⎞
⎠
⎟ = +
1
p
z
K z
b
b
=
+
31. 18 Textbook of Receptor Pharmacology, Second Edition
muscle fibers. So, when the twitch of the whole muscle has fallen to half its initial amplitude,
receptor occupancy by tubocurarine in the surface fibers is much greater than 50%.
The second complication is that the rate at which a ligand acts on an isolated tissue is often
determined by the diffusion of ligand molecules through the tissue rather than by their combination
with the receptors.Again taking as our example the action of tubocurarine on the isolated diaphragm,
the slow development of the block reflects not the rate of binding to the receptors but rather the
failure of neuromuscular transmission in an increasing number of individual muscle fibers as
tubocurarine slowly diffuses between the closely packed fibers into the interior of the preparation.
Moreover, as an individual ligand molecule passes deeper into the tissue, it may bind and unbind
several times (and for different periods) to a variety of sites (including receptors). This repeated
binding and dissociation can greatly slow diffusion into and out of the tissue.
For these reasons, kinetic measurements are now usually done with isolated cells (e.g., a single
neuron or a muscle fiber) or even a patch of cell membrane held on the tip of a suitable microelec-
trode. Another approach is to work with a cell membrane preparation and examine directly the rate
at which a suitable radioligand combines with, or dissociates from, the receptors that the membrane
carries. Our next task is to consider what binding kinetics might be expected under such conditions.
1.3.2 INCREASES IN RECEPTOR OCCUPANCY
In the following discussion, we continue with the simple model for the combination of a ligand
with its binding sites that was introduced in Section 1.2.1 (Eq. (1.1)). Assuming as before that the
law of mass action applies, the rate at which receptor occupancy (pAR) changes with time should
be given by the expression:
(1.18)
In words, this states that the rate of change of occupancy is simply the difference between the rate
at which ligand–receptor complexes are formed and the rate at which they break down.*
At first sight, Eq. (1.18) looks difficult to solve because there are no less than four variables:
pAR, t, [A], and pR. However, we know that pR = (1 – pAR). Also, we will assume, as before, that
[A] remains constant; that is, so much A is present in relation to the number of binding sites that
the combination of some of it with the sites will not appreciably reduce the overall concentration.
Hence, only pAR and t remain as variables, and the equation becomes easier to handle.
Substituting for pR, we have:
(1.19)
Rearranging terms,
(1.20)
This still looks rather complicated, so we will drop the subscript from pAR and make the following
substitutions for the constants in the equation:
* If the reader is new to calculus or not at ease with it, a slim volume (Calculus Made Easy) by Silvanus P. Thompson is
strongly recommended.
d
d
A
AR
R AR
( )
[ ]
p
t
k p k p
= −
+ −
1 1
d
d
A
AR
AR AR
( )
[ ]( )
p
t
k p k p
= − −
+ −
1 1
1
d
d
A A
AR
AR
( )
[ ] ( [ ])
p
t
k k k p
= − +
+ − +
1 1 1
32. Classical Approaches to the Study of Drug–Receptor Interactions 19
a = k+1[A]
b = k–1 + k+1[A]
Hence,
This can be rearranged to a standard form that is easily integrated to determine how the occupancy
changes with time:
Integrating,
We can now consider how quickly occupancy rises after the ligand is first applied, at time zero
(t1 = 0). Receptor occupancy is initially 0, so that p1 is 0. Thereafter, occupancy increases steadily
and will be denoted pAR(t) at time t:
Hence,
Replacing a and b by the original terms, we have:
(1.21)
Recalling that k–1/k+1 = KA, we can write:
t1 = 0 p1 = 0
t2 = t p2 = pAR(t)
d
d
p
t
a bp
= −
d
d
p
a bp
t
t
t
p
p
−
=
∫
∫ 1
2
1
2
log ( )
e
a bp
a bp
b t t
−
−
⎛
⎝
⎜
⎞
⎠
⎟ = − −
2
1
2 1
log
( )
( )
( ) ( )
e
bt
bt
a bp t
a
bt
a bp t
a
e
p t
a
b
e
−
⎧
⎨
⎩
⎫
⎬
⎭
= −
−
=
= −
−
−
AR
AR
AR 1
p t
k
k k
e k k t
AR
A
A
A
( )
[ ]
[ ]
( [ ])
=
+
−
{ }
+
− +
− +
− +
1
1 1
1 1 1
p t
K
e k k t
AR
A
A
A
A
( )
[ ]
[ ]
( [ ])
=
+
−
{ }
− +
− +
1 1 1
33. 20 Textbook of Receptor Pharmacology, Second Edition
When t is very great, the ligand and its binding sites come into equilibrium. The term in large
brackets then becomes unity (because e–∞ = 0) so that
We can then write:
(1.22)
This is the expression we need. It has been plotted in Figure 1.3 for three concentrations of A.
Note how the rate of approach to equilibrium increases as [A] becomes greater. This is because
the time course is determined by (k–1 + k+1[A]). This quantity is sometimes replaced by a single
constant, so that Eq. (1.22) can be rewritten as either:
pAR(t) = pAR(∞)(1 – e–λt) (1.23)
or
pAR(t) = pAR(∞)(1 – e–t/τ) (1.24)
where
λ = k–1 + k+1[A] = 1/τ
where τ (tau) is the time constant and has the unit of time; λ (lambda) is the rate constant, which
is sometimes written as kon (as in Chapter 5) and has the unit of time–1.
FIGURE 1.3 The predicted time course of the rise in receptor occupancy following the application of a ligand
at the three concentrations shown. The curves have been drawn according to Eq. (1.22), using a value of 2 ×
106 M–1sec–1 for k+1 and of 1 sec–1 for k–1.
p
K
AR
A
A
A
( )
[ ]
[ ]
∞ =
+
p t p e k k t
AR AR
A])
( ) ( ){ }
( [
= ∞ − − +
− +
1 1 1
34. Classical Approaches to the Study of Drug–Receptor Interactions 21
1.3.3 FALLS IN RECEPTOR OCCUPANCY
Earlier, we had assumed for simplicity that the occupancy was zero when the ligand was first
applied. It is straightforward to extend the derivation to predict how the occupancy will change
with time even if it is not initially zero. We alter the limits of integration to
Here, pAR(0) is the occupancy at time zero, and the other terms are as previously defined.
Exactly the same steps as before then lead to the following expression to replace Eq. (1.22):
(1.25)
We can use this to examine what would happen if the ligand is rapidly removed. This is
equivalent to setting [A] abruptly to zero, at time zero, and p(∞) also becomes zero because
eventually all the ligand receptor complexes will dissociate. Eq. (1.25) then reduces to:
(1.26)
This expression has been plotted in Figure 1.4.
The time constant, τ, for the decline in occupancy is simply the reciprocal of k–1. A related
term is the half-time (t1/2). This is the time needed for the quantity (pAR(t) in this example) to reach
halfway between the initial and the final value and is given by:
For the example illustrated in Figure 1.4, t1/2 = 0.693 sec. Note that τ and t1/2 have the unit of time,
as compared with time–1 for k–1.
FIGURE 1.4 The predicted time course of the decline in binding-site occupancy. The lines have been plotted
using Eq. (1.26), taking k–1 to be 1 sec–1 and pAR(0) to be 0.8. A linear scale for pAR(t) has been used on the
left, and a logarithmic one on the right.
t1 = 0 p1 = pAR(0)
t2 = t p2 = pAR(t)
p t p p p e k k t
AR AR AR AR
A
( ) ( ) { ( ) ( )} ( [ ])
= ∞ + − ∞ − +
− +
0 1 1
p t p e k t
AR AR
( ) ( )
= − −
0 1
t
k
1 2
1
0 693
/
.
=
−
35. 22 Textbook of Receptor Pharmacology, Second Edition
It has been assumed in this introductory account that so many binding sites are present that
the average number occupied will rise or fall smoothly with time after a change in ligand
concentration; events at single sites have not been considered. When a ligand is abruptly removed,
the period for which an individual binding site remains occupied will, of course, vary from site
to site, just as do the lifetimes of individual atoms in a sample of an element subject to radioactive
decay. It can be shown that the median lifetime of the occupancy of individual sites is given by
0.693/k–1. The mean lifetime is 1/k–1. The introduction of the single-channel recording method
has made it possible to obtain direct evidence about the duration of receptor occupancy (see
Chapter 6).
1.4 PARTIAL AGONISTS
1.4.1 INTRODUCTION AND EARLY CONCEPTS
The development of new drugs usually requires the synthesis of large numbers of structurally related
compounds. If a set of agonists of this kind is tested on a particular tissue, the compounds are often
found to fall into two categories. Some can elicit a maximal tissue response and are described as
full agonists in that experimental situation. Others cannot elicit this maximal response, no matter
how high their concentration, and are termed partial agonists. Examples include:
Figure 1.5 shows concentration–response curves that compare the action of the β-adrenoceptor
partial agonist prenalterol with that of the full agonist isoprenaline on a range of tissues and
responses. In every instance, the maximal response to prenalterol is smaller, though the magnitude
of the difference varies greatly.
It might be argued that a partial agonist cannot match the response to a full agonist because it
fails to combine with all the receptors. This can easily be ruled out by testing the effect of increasing
concentrations of a partial agonist on the response of a tissue to a fixed concentration of a full
agonist. Figure 1.6 (right, upper curve) illustrates such an experiment for two agonists acting at H2
receptors. As the concentration of the partial agonist impromidine is raised, the response of the
tissue gradually falls from the large value seen with the full agonist alone and eventually reaches
the maximal response to the partial agonist acting on its own. The implication is that the partial
agonist is perfectly capable of combining with all the receptors, provided that a high enough
concentration is applied, but the effect on the tissue is less than what would be seen with a full
agonist. The partial agonist is in some way less able to elicit a response.
The experiment of Figure 1.7 points to the same conclusion. When very low concentrations of
histamine are applied in the presence of a relatively large fixed concentration of impromidine, the
overall response is mainly due to the receptors occupied by impromidine; however, the concentra-
tion–response curves cross as the histamine concentration is increased. This is because the presence
of impromidine reduces receptor occupancy by histamine (at all concentrations) and vice versa.
When the lines intersect, the effect of the reduction in impromidine occupancy by histamine is
exactly offset by the contribution from the receptors occupied by histamine. Beyond this point, the
presence of impromidine lowers the response to a given concentration of histamine. In effect, it
acts as an antagonist. Again, the implication is that the partial agonist can combine with all the
receptors but is less able to produce a response.
Partial Agonist Full Agonist Acting at:
Prenalterol Adrenaline, isoprenaline β-Adrenoceptors
Pilocarpine Acetylcholine Muscarinic receptors
Impromidine Histamine Histamine H2 receptors
36. Classical Approaches to the Study of Drug–Receptor Interactions 23
FIGURE 1.5 Comparison of the log concentration–response relationships for β-adrenoceptor-mediated
actions on six tissues of a full and a partial agonist (isoprenaline [closed circles] and prenalterol [open circles],
respectively). The ordinate shows the response as a fraction of the maximal response to isoprenaline. (From
Kenakin, T. P. and Beek, D., J. Pharmacol. Exp. Ther., 213, 406–413, 1980.)
FIGURE 1.6 Interaction between the full agonist histamine and the H2-receptor partial agonist impromidine
on isolated ventricular strips from human myocardium. The concentration–response curve on the left is for
histamine alone, and those on the right show the response to impromidine acting either on its own (open
squares) or in the presence of a constant concentration (100 µM) of histamine (open diamonds). (From English,
T. A. H. et al., Br. J. Pharmacol., 89, 335–340, 1986.)
37. 24 Textbook of Receptor Pharmacology, Second Edition
1.4.2 EXPRESSING THE MAXIMAL RESPONSE TO A PARTIAL AGONIST: INTRINSIC
ACTIVITY AND EFFICACY
In 1954 the Dutch pharmacologist E. J. Ariëns introduced the term intrinsic activity, which is now
usually defined as:
For full agonists, the intrinsic activity (often denoted by α) is unity, by definition, as compared
with zero for a competitive antagonist. Partial agonists have values between these limits. Note that
the definition is entirely descriptive; nothing is assumed about mechanism. Also, intrinsic should
not be taken to mean that a given agonist has a characteristic activity, regardless of the experimental
circumstances. To the contrary, the intrinsic activity of a partial agonist such as prenalterol can
vary greatly not only between tissues, as Figure 1.5 illustrates, but also in a given tissue, depending
on the experimental conditions (see later discussion). Indeed, the same compound can be a full
agonist with one tissue and a partial agonist with another. For this reason, the term maximal agonist
effect is perhaps preferable to intrinsic activity.
FIGURE 1.7 Log concentration–response curves for histamine applied alone (open circles) or in the presence
(open squares) of a constant concentration of the partial agonist impromidine (10 µM). Tissue and experimental
conditions as in Figure 1.6. (From English, T. A. H. et al., Br. J. Pharmacol., 89, 335–340, 1986.)
Intrinsic activity =
maximum response to test agonist
maximum response to a full agonist acting through the same receptors
38. Classical Approaches to the Study of Drug–Receptor Interactions 25
Similarly, the finding that a pair of agonists can each elicit the maximal response of a tissue
(i.e., they have the same intrinsic activity, unity) should not be taken to imply that they are equally
able to activate receptors. Suppose that the tissue has many spare receptors (see Section 1.6.3).
One of the agonists might have to occupy 5% of the receptors in order to produce the maximal
response, whereas the other might require only 1% occupancy. Evidently, the second agonist is
more effective, despite both being full agonists. A more subtle measure of the ability of an agonist
to activate receptors is clearly necessary, and one was provided by R. P. Stephenson, who suggested
that receptor activation resulted in a “stimulus” or “signal” (S) being communicated to the cells,
and that the magnitude of this stimulus was determined by the product of what he termed the
efficacy (e) of the agonist and the proportion, p, of the receptors that it occupies:*
S = ep (1.27)
An important difference from Ariëns’s concept of intrinsic activity is that efficacy, unlike
intrinsic activity, has no upper limit; it is always possible that an agonist with a greater efficacy
than any existing compound may be discovered. Also, Stephenson’s proposal was not linked to any
specific assumption about the relationship between receptor occupancy and the response of the
tissue. (Ariëns, like A. J. Clark, had initially supposed direct proportionality, an assumption later
to be abandoned.) According to Stephenson,
(1.28)
Here, y is the response of the tissue, and eA is the efficacy of the agonist A. f(SA) means merely
“some function of SA” (i.e., y depends on SA in some as yet unspecified way). Note that, in keeping
with the thinking at the time, Stephenson used the Hill–Langmuir equation to relate agonist
concentration, [A], to receptor occupancy, pAR. This most important assumption is reconsidered in
the next section.
In order to be able to compare the efficacies of different agonists acting through the same
receptors, Stephenson proposed the convention that the stimulus S is unity for a response that is
50% of the maximum attainable with a full agonist. This is the same as postulating that a partial
agonist that must occupy all the receptors to produce a half-maximal response has an efficacy of
unity. We can see this from Eq. (1.27); if our hypothetical partial agonist has to occupy all the
receptors (i.e., p = 1) to produce the half-maximal response, at which point S also is unity (by
Stephenson’s convention), then e must also be 1.
R. F. Furchgott later suggested a refinement of Stephenson’s concept. Recognizing that the
response of a tissue to an agonist is influenced by the number of receptors as well as by the ability
of the agonist to activate them, he wrote:
e = ε[R]T
Here, [R]T is the total “concentration” of receptors, and ε (epsilon) is the intrinsic efficacy (not to
be confused with intrinsic activity); ε can be regarded as a measure of the contribution of individual
receptors to the overall efficacy.
The efficacy of a particular agonist, as defined by Stephenson, can vary between different tissues
in the same way as can the intrinsic activity, and for the same reasons. Moreover, the value of both
the intrinsic activity and the efficacy of an agonist in a given tissue will depend on the experimental
* No distinction is made here between occupied and activated receptors. This point is of key importance, as already noted
in Section 1.2.3, and is discussed further in the following pages.
y f S f e p f e
K
= = =
+
⎛
⎝
⎜
⎞
⎠
⎟
( ) ( )
[ ]
[ ]
A A AR A
A
A
A
39. 26 Textbook of Receptor Pharmacology, Second Edition
conditions, as illustrated in Figure 1.8. Relaxations of tracheal muscle in response to the β-adreno-
ceptor agonists isoprenaline and prenalterol were measured first in the absence (circles) and then in
the presence (triangles, squares) of a muscarinic agonist, carbachol, which causes contraction and so
tends to oppose β-adrenoceptor-mediated relaxation. Hence, greater concentrations of the β-agonists
are needed, and the curves shift to the right. With isoprenaline, the maximal response can still be
obtained, despite the presence of carbachol at either concentration. The pattern is quite different with
prenalterol. Its inability to produce complete relaxation becomes even more evident in the presence
of carbachol at 1 µM. Indeed, when administered with 10 µM carbachol, prenalterol causes little or
no relaxation; its intrinsic activity and efficacy (in Stephenson’s usage) have become negligible.
In the same way, reducing the number of available receptors (for example, by applying an
alkylating agent; see Section 1.6.1) will always diminish the maximal response to a partial agonist.
In contrast, the log concentration–response curve for a full agonist may first shift to the right, and
the maximal response will become smaller only when no spare receptors are available for that
agonist (see Section 1.6.3). Conversely, increasing the number of receptors (e.g., by upregulation
or by deliberate overexpression of the gene coding for the receptor) will cause the maximal response
to a partial agonist to become greater, whereas the log concentration–response curve for a full
agonist will move to the left.
1.4.3 INTERPRETATION OF PARTIAL AGONISM IN TERMS OF EVENTS AT INDIVIDUAL
RECEPTORS
The concepts of intrinsic activity and efficacy just outlined are purely descriptive, without reference
to mechanism. We turn now to how differences in efficacy might be explained in terms of the
molecular events that underlie receptor activation, and we begin by considering some of the
experimental evidence that has provided remarkably direct evidence of the nature of these events.
Just a year after Stephenson’s classical paper of 1956, J. del Castillo and B. Katz published an
electrophysiological study of the interactions that occurred when pairs of agonists with related
structures were applied simultaneously to the nicotinic receptors at the endplate region of skeletal
muscle. Their findings could be best explained in terms of a model for receptor activation that has
already been briefly introduced in Section 1.2.3 (see particularly Eq. (1.7)). In this scheme, the
occupied receptor can isomerize between an active and an inactive state. This is very different from
the classical model of Hill, Clark, and Gaddum in which no clear distinction was made between
the occupation and activation of a receptor by an agonist.
FIGURE 1.8 The effect of carbachol at two concentrations, 1 µM (triangles) and 10 µM (squares), on the
relaxations of tracheal smooth muscle caused by a partial agonist, prenalterol, and by a full agonist, isopre-
naline. The responses are plotted as a fraction of the maximum to isoprenaline. (From Kenakin, T. P. and
Beek, D., J. Pharmacol. Exp. Ther., 213, 406–413, 1980.)
40. Classical Approaches to the Study of Drug–Receptor Interactions 27
Direct evidence for this action was to come from the introduction by E. Neher and B. Sakmann
in 1976 of the single-channel recording technique, which allowed the minute electrical currents
passing through the ion channel intrinsic to the nicotinic receptor, and other ligand-gated ion
channels, to be measured directly and as they occurred. For the first time it became possible to
study the activity of individual receptors in situ (see also Chapter 6). It was quickly shown that for
a wide range of nicotinic agonists, these currents had exactly the same amplitude. This is illustrated
for four such agonists in Figure 1.9. What differed among agonists was the fraction of time for
which the current flowed (i.e., for which the channels were open). This is just what would be
expected from the del Castillo–Katz scheme if the active state (AR*) of the occupied receptor is
the same (in terms of the flow of ions through the open channel) for different agonists. However,
with a weak partial agonist, the receptor is in the AR* state for only a small fraction of the time,
even if all the binding sites are occupied.
FIGURE 1.9 Records of the minute electrical currents (downward deflections) that flow through single ligand-
gated ion channels in the junctional region of frog skeletal muscle. The currents arise from brief transitions
of individual nicotinic receptors to an active (channel open) state in response to the presence of various
agonists (ACh = acetylcholine; SubCh = suberyldicholine; DecCh = the dicholine ester of decan-1,10-
dicarboxylic acid; CCh = carbamylcholine). (From Colquhoun, D. and Sakmann, B., J. Physiol., 369, 501–557,
1985. With permission.)
41. 28 Textbook of Receptor Pharmacology, Second Edition
The next question to consider is the interpretation of efficacy (both in the particular sense
introduced by Stephenson and in more general terms) in the context of the model proposed by del
Castillo and Katz.
1.4.4 THE DEL CASTILLO–KATZ MECHANISM: 1. RELATIONSHIP BETWEEN AGONIST
CONCENTRATION AND FRACTION OF RECEPTORS IN AN ACTIVE FORM
Our first task is to apply the law of mass action to derive a relationship between the concentration
of agonist and the proportion of receptors that are in the active form at equilibrium. This proportion
will be denoted by pAR*.
As in all the derivations in this chapter, this one requires only three steps. The first is to apply
the law of mass action to each of the equilibria that exist. The second is to write an equation that
expresses the fact that the fractions of receptors in each condition that can be distinguished must
add up to 1 (the “conservation rule”). The del Castillo–Katz scheme in its simplest form (see Eq.
(1.7) in Section 1.2.3) has three such conditions: R (vacant and inactive), AR (inactive, though A
is bound), and AR* (bound and active). The corresponding fractions of receptors in these condi-
tions* are pR, pAR, and pAR*.
Applying the law of mass action to each of the two equilibria gives:
[A]pR = KA pAR (1.29)
pAR* = E pAR (1.30)
where KA and E are the equilibrium constants indicated in Eq. (1.7).
Also,
pR + pAR + pAR* = 1 (1.31)
We can now take the third and last step. What we wish to know is pAR*, so we use Eqs. (1.29)
and (1.30) to substitute for pR and pAR in Eq. (1.31), obtaining:
(1.32)
This is the expression we require. Although it has the same general form as the Hill–Langmuir
equation, two important differences are to be noted:
1. As [A] is increased, pAR* tends not to unity but to
* The term “state” rather than “condition” is often used in this context. However, the latter seems preferable in an introductory
account. This is because the del Castillo–Katz mechanism is often described as a “two-state” model of receptor action,
meaning here that the occupied receptor exists in two distinct (albeit interconvertible) forms, AR and AR*, whereas three
conditions of the receptor (R, AR, and AR*) have to be identified when applying the law of mass action to the binding of
the ligand, A.
K
E
p
E
p p
A
AR AR AR*
A
[ ] * *
+ + =
1
1
∴ =
+ +
p
E
K E
AR*
A
A
A]
[ ]
( )[
1
E
E
1+
42. Classical Approaches to the Study of Drug–Receptor Interactions 29
Thus, the value of E will determine the maximal response to A. Only if E is very large
in relation to one will almost all the receptors be activated, as illustrated in Figure 1.10,
which plots Eq. (1.32) for a range of values of E.
2. Equation (1.32) gives the proportion of active receptors (pAR*), rather than occupied
receptors (pocc = pAR + pAR*). To obtain the occupancy, we can use Eq. (1.30) to express
pAR in terms of pAR*:
(1.33)
(1.34)
This can be rewritten as:
(1.35)
where Keff, the effective dissociation equilibrium constant, is defined as:
(1.36)
Because Keff applies to a scheme that involves more that one equilibrium (see Eq. (1.7)),
it is referred to as a macroscopic equilibrium constant, to distinguish it from the micro-
scopic equilibrium constants KA and E, which describe the individual equilibria.
FIGURE 1.10 The relationship between pAR* and [A] predicted by Eq. (1.32) for a range of values of E (given
with each line). Note that as E rises above 10, the curves move to the left even though the value of KA, the
dissociation equilibrium constant for the initial combination ofA with its binding site, is 200 µM for each curve.
p p p
E
E
p
E
K E
occ AR AR* AR*
A
A
A]
= + =
+
⎛
⎝
⎞
⎠
=
+
+ +
1
1
1
( )[ ]
( )[
=
+
+
[ ]
[
A
A]
A
K
E
1
p
K
occ
eff
A
A]
=
+
[ ]
[
K
K
E
eff
A
=
+
1
43. 30 Textbook of Receptor Pharmacology, Second Edition
These results show that if the relationship between the concentration of an agonist and the
proportion of receptors that it occupies is measured directly (e.g., using a radioligand binding
method), the outcome should be a simple hyperbolic curve. Although the curve is describable by
the Hill–Langmuir equation, the dissociation equilibrium constant for the binding will be not KA
but Keff, which is determined by both E and KA.
1.4.5 THE DEL CASTILLO–KATZ MECHANISM: 2. INTERPRETATION OF EFFICACY FOR
LIGAND-GATED ION CHANNELS
In general terms, it is easy to see that the value of the equilibrium constant E in Eq. (1.7) will
determine whether a ligand is a full agonist, a partial agonist, or an antagonist. We first recall
Stephenson’s concept that the response of a tissue to an agonist is determined by the product, S, of
the efficacy of the agonist and the proportion of receptors occupied (see Eq. (1.27)). To relate this
to the del Castillo–Katz scheme, we rewrite Eq. (1.33) to show the relation between the proportion
of active receptors, pAR* (which determines the tissue response) and total receptor occupancy:
(1.37)
From this we can see that the term E/(1 + E) is equivalent, in a formal sense at least, to
Stephenson’s efficacy. If an agonist is applied at a very high concentration, so that all the receptors
are occupied, the proportion in the active state is E/(1 + E). If this agonist is also very effective
(i.e., if E is >>1), the proportion of active receptors becomes close to unity, the upper limit.
Consider next a hypothetical partial agonist that, even when occupying all the receptors (pocc = 1),
causes only half of them to be in the active form (i.e., pAR = pAR* = 0.5). From Eq. (1.37), we can
see that E must be unity for this agonist. In Stephenson’s scheme, such an agonist would have an
efficacy of unity, provided that the response measured is a direct indication of the proportion of
activated receptors.
The realization that the ability of an agonist to activate a receptor can be expressed in this way
has led to great interest in measuring the rate constants (two each for KA and E, at the simplest)
that determine not only the values of KA and E but also the kinetics of agonist action. The single-
channel recording technique allows this to be achieved for ligand-gated ion channels, as described
in Chapter 6. Note, however, the complication that such receptors generally carry two binding sites
for the agonist, so the simple scheme just considered, Eq. (1.7), has to be elaborated (see Eq. (1.9)
in Appendix 1.2C [Section 1.2.4.3] and also Chapter 6).
A difficulty encountered in such work, and one that has to be considered in any study of the
relationship between the concentration of an agonist and its action, is the occurrence of desensiti-
zation. The response declines despite the continued presence of the agonist. Several factors can
contribute. One that has been identified in work with ligand-gated ion channels is that receptors
occupied by agonist and in the active state (AR*) may isomerize to an inactive, desensitized, state,
ARD. This can be represented as:
As explained in Chapter 6, quantitative studies of desensitization at ligand-gated ion channels
have shown that even this scheme is an oversimplification, and it is necessary to include the
possibility that receptors without ligands can exist in a desensitized state.
p
E
E
p
AR* occ
=
+
1
A R AR AR AR
inactive inactive active
D
inactive
+
( ) ( ) ( ) ( )
*
K E K
A D
44. Classical Approaches to the Study of Drug–Receptor Interactions 31
Desensitization can occur in other ways. With G-protein-coupled receptors, it can result from
phosphorylation of the receptor by one or more protein kinases that become active following the
application of agonist.* This activation is sometimes followed by the loss of receptors from the
cell surface. An agonist-induced reduction in the number of functional receptors over a relatively
long time period is described as downregulation. Receptor upregulation can also occur, for example,
following the prolonged administration of antagonists in vivo.
1.4.6 INTERPRETATION OF EFFICACY FOR RECEPTORS ACTING THROUGH
G-PROTEINS
Some of the most revealing studies of partial agonism (including Stephenson’s seminal work) have
been done with tissues in which G-proteins (see Chapters 2 and 7) provide the link between receptor
activation and initiation of the response. In contrast to the situation with “fast” receptors with
intrinsic ion channels (see above), it is not yet possible to observe the activity of individual G-
protein-coupled receptors (with the potential exception of some that are linked to potassium
channels); however, enough is known to show that the mechanisms are complex. The interpretation
of differences in efficacy for agonists acting at such receptors is correspondingly less certain.
An early model for the action of such receptors was as follows:
Here, the agonist–receptor complex (AR) combines with a G-protein (G) to form a ternary complex
(ARG*), which can initiate further cellular events, such as the activation of adenylate cyclase.
However, this simple scheme (the ternary complex model) was not in keeping with what was already
known about the importance of isomerization in receptor activation (see Sections 1.2.3 and 1.4.3),
and it also failed to account for findings that were soon to come from studies of mutated receptors.
In all current models of G-protein-coupled receptors, receptor activation by isomerization is
assumed to occur so that the model becomes:
(1.38)
Here, combination of the activated receptor (AR*) with the G-protein causes the latter to enter an
active state (G*) which can initiate a tissue response through, for example, adenylate cyclase,
phospholipase C, or the opening or closing of ion channels. In this scheme, what will determine
whether a particular agonist can produce a full or only a limited response? Suppose that a high
concentration of the agonist is applied, so that all the receptors are occupied. They will then be
distributed among the AR, AR*, and AR*G* conditions, of which AR*G* alone leads to a response.
The values of both E and KARG will then influence how much AR*G* is formed, and hence whether
the agonist in question is partial or otherwise. In principle, each of these two equilibrium constants
* Some of these protein kinases are specific for particular receptors (e.g., β-adrenergic receptor kinase [βARK], now referred
to as GRK2).
A R AR
AR G ARG
A
ARG
+
+ *
K
K
A R AR AR
AR G AR G
A
ARG
+
+
*
* * *
K E
K
45. 32 Textbook of Receptor Pharmacology, Second Edition
could vary from agonist to agonist. By analogy with ligand-gated ion channels, it is tempting to
suppose that only E is agonist dependent and that the affinity of the active, AR*, state of the receptor
for the G-protein is the same for all agonists. However, in the absence of direct evidence, this must
remain an open question. Note that, in any case, the magnitude of the response may also depend
on the availability of the G-protein. If very little is available, only a correspondingly small amount
of AR*G* can be formed, regardless of the concentration of agonist and the number of receptors.
Similarly, if few receptors are present in relation to the total quantity of G-protein, that too will
limit the formation of AR*G*. Thus, the maximum response to an agonist is influenced by tissue
factors as well as by KA, E, and KARG. This can be shown more formally by applying the law of
mass action to the three equilibria shown in Eq. (1.38). The outcome, with some further discussion,
is given in Appendix 1.4B (Section 1.4.9.2).
Complicated though these schemes might seem, they are in fact oversimplifications. Factors
that have not been considered include:
1. It is likely that some receptors are coupled to G-proteins even in the absence of agonist.
2. The activated receptor combines with the G-protein in its GGDP form, with the conse-
quence that guanosine triphosphate (GTP) can replace previously bound guanosine
diphosphate (GDP). The extent to which this can occur will be influenced by the local
concentration of GTP.
3. The structure of the G-protein is heterotrimeric. Following activation by GTP binding,
the trimer dissociates into its α and βγ subunits, each of which may elicit cell responses.
4. G-protein activation has a cyclical nature. The α subunit can hydrolyze the GTP that is
bound to it, thereby allowing the heterotrimer to reform. The lifetime of individual αGTP
subunits will vary (cf. the lifetimes of open ion channels).
5. More than one type of G-protein, each with characteristic cellular actions, may be present
in many cells.
6. Some G-protein-coupled receptors have been found to be constitutively active (see the
following section).
7. It is possible (although as yet unproven) that the affinity of the active form of the occupied
receptor (AR*) for the G-protein may vary from agonist to agonist.
8. Recent evidence suggests that several G-protein-coupled receptors exist as dimers.
In principle, these features can be built into models of receptor activation, although the large
number of disposable parameters makes testing difficult. Some of the rate and equilibrium constants
must be known beforehand. One experimental tactic is to alter the relative proportions of receptors
and G-protein and then determine whether the efficacy of agonists changes in the way expected
from the model. The discovery that some receptors are constitutively active has provided another
new approach as well as additional information about receptor function, as we shall now see.
1.4.7 CONSTITUTIVELY ACTIVE RECEPTORS AND INVERSE AGONISTS
The del Castillo–Katz scheme (in common, of course, with the simpler model explored by Hill,
Clark, and Gaddum) supposes that the receptors are inactive in the absence of agonist. It is now
known that this is not always so; several types of receptor are constitutively active. Examples
include mutated receptors responsible for several genetically determined diseases. Thus, hyperthy-
roidism can result from mutations that cause the receptors for thyrotropin (TSH, or thyroid-
stimulating hormone) to be active even in the absence of the hormone. Also, receptor variants that
are constitutively active have been created in the laboratory by site-directed mutagenesis. Finally,
deliberate overexpression of receptors by receptor-gene transfection of cell lines and even laboratory
animals has revealed that many “wild-type” receptors also show some activity in the absence of
agonist. What might the mechanism be? The most likely possibility, and one which is in keeping
46. Classical Approaches to the Study of Drug–Receptor Interactions 33
with what has been learned about how ion channels work, is that such receptors can isomerize
spontaneously to and from an active form:
In principle, both forms could combine with agonist, or indeed with any ligand, L, with affinity,
as illustrated in Figure 1.11.
Suppose that L combines only with the inactive, R, form. Then the presence of L, by promoting
the formation of LR at the expense of the other species, will reduce the proportion of receptors in
the active, R*, state. L is said to be an inverse agonist or negative antagonist and to possess negative
efficacy. If, in contrast, L combines with the R* form alone, it will act as a conventional or positive
agonist of very high intrinsic efficacy.
Exploring the scheme further, a partial agonist will bind to both R and R* but with some
preferential affinity for one or the other of the two states. If the preference is for R, the ligand will
be a partial inverse agonist, as its presence will reduce the number of receptors in the active state,
though not to zero.
As shown in Section 1.10 (see the solution to Problem 1.4), application of the law of mass
action to the scheme of Figure 1.11 provides the following expression for the fraction of receptors
in the active state (i.e., pR* + pLR*) at equilibrium:
(1.39)
Here, the equilibrium constant E0 is defined by pR*/pR, KL by [L]pR/pLR, and by [L]pR*/pLR*.
Figure 1.12 plots this relationship for three hypothetical ligands that differ in their relative affinities
for the active and the inactive states of the receptor. The term α has been used to express the ratio
of KL to . When α = 0.1, the ligand is an inverse agonist; whereas when α = 100, it is a
conventional agonist. In the third example, with a ligand that shows no selectivity between the
active and inactive forms of the receptor (α = 1), the proportion of active receptors remains
unchanged as [L] (and therefore receptor occupancy) is increased.
Such a ligand, however, will reduce the action of either a conventional or an inverse agonist,
and so in effect is an antagonist. More precisely, it is a neutral competitive antagonist. If large
FIGURE 1.11 A model to show the influence of a ligand, L, on the equilibrium between the active and
inactive forms of a constitutively active receptor, R. Note that if L, R, and LR are in equilibrium, and likewise
L, R* and LR*, then the same must hold for LR and LR* (see Appendix 1.6B (Section 1.6.7.2) for further
explanation).
R R
inactive active
( ) ( )
*
p
E
E
K
K
active
L
L
*
L
1+
[L]
=
+
+
⎛
⎝
⎜
⎜
⎜
⎞
⎠
⎟
⎟
⎟
0
0
1
[ ]
KL
*
KL
*
⎟
⎜
47. 34 Textbook of Receptor Pharmacology, Second Edition
numbers of competitive antagonists of the same pharmacological class (e.g., β-adrenoceptor block-
ers) are carefully tested on a tissue or cell line showing constitutive activity, some will be found
to cause a small increase in basal activity. They are, in effect, weak conventional partial agonists.
Others will reduce the basal activity and so may be inverse agonists with what could be a substantial
degree of negative efficacy.* Few of the set can be expected to have exactly the same affinity for
the active and inactive forms of the receptor and so be neutral antagonists. However, some com-
pounds of this kind have been identified, and Figure 1.13 illustrates the effect of one on the response
to both a conventional and an inverse agonist acting on 5HT1A receptors expressed in a cell line.
As with the experiments of Figure 1.13, constitutive activity is often investigated in cultured
cell lines that do not normally express the receptor to be examined but have been made to do so
by transfection with the gene coding for either the native receptor or a mutated variant of it. The
number of receptors per cell (receptor density) may be much greater in these circumstances than
in cells that express the receptors naturally. While overexpression of this kind has the great advantage
that small degrees of constitutive activity can be detected and studied, it is worth noting that
constitutive activity is often much less striking in situ than in transfected cells. Hence, the partial
agonist action (conventional or inverse) of an antagonist may be much less marked, or even
negligible, when studied in an intact tissue so that simple competitive antagonism is observed, as
described in Section 1.5.
Nevertheless, the evidence that some receptors have sufficient constitutive activity to influence
cell function in vivo even in the absence of agonist makes it necessary to extend the simple models
already considered for the activation of G-protein-coupled receptors. In principle, the receptor can
now exist in no less than eight different conditions (R, R*, LR, LR*, RG, R*G, LRG, LR*G),
which is best represented graphically as a cube with one of the conditions at each vertex (see Figure
1.14). The calculation of the proportions of activated and occupied receptors is straightforward, if
lengthy (see the answer to Problem 1.5 in Section 1.10). Finding the proportion in the active form
is more difficult if the supply of G-protein is limited but can be done using numerical methods.
FIGURE 1.12 The relationship between the total fraction of receptors in the active state (pR* + pAR*) and
ligand concentration ([L]) for a constitutively active receptor. The curve has been drawn according to Eq.
(1.39), using the following values: E0 = 0.2, KL = 200 nM, α = KL/ = 0.1, 1, and 100, as shown. Note that
on this model some of the receptors (a fraction given by E0/(1 + E0) = 0.167) are active in the absence of ligand.
* The possibility that the depression in basal activity may have some other explanation (e.g., an inhibitory action on one
or more of the events that follow receptor activation) should not be overlooked.
KL
*
48. Classical
Approaches
to
the
Study
of
Drug–Receptor
Interactions
35
FIGURE 1.13 The effects of a conventional agonist, an inverse agonist, and a neutral antagonist on the activity of a constitutively active G-protein-coupled 5HT1A
receptor. The panel on the left shows the log concentration–response curves for the conventional agonist (open squares) and the inverse agonist (open circles). The closed
symbols show how the curves change when the antagonist (WAY 100,635 at 10 nM) is included in the incubation fluid. Note the parallel, and similar, shift in the lines.
The panel on the right illustrates the effects of a wide range of concentrations of the same antagonist applied on its own (open diamonds) or in the presence of a high
concentration of either the conventional agonist (closed squares) or the inverse agonist (closed circles). Note that the antagonist by itself causes little change, showing
that it has no preference for the active or inactive forms of the receptor. In keeping with this, high concentrations of the antagonist abolish the response to both types of
agonist (the curves converge). (From Newman-Tancredi, A. et al., Br. J. Pharmacol., 120, 737–739, 1997.)